Generalized Multi-Release Framework for Fault Prediction in Open Source Software

Literature Review

In the field of information technology, reliable software has become the most important concern. Software reliability is defined as “the probability of error free operation of the software for a given time period within some specified environment” (Kapur et al., 2011; Pham, 2006). Software Reliability Growth Models are used to quantitatively estimate the reliability of the software during the testing phase. A mathematical relationship between the testing time and total number of faults is given by SRGMs. Although a number of reliability models have been proposed so far, but all these are based on some limitations (Musa, 2004; Kapur et al., 2011).

Tremendous work has been done to assess the reliability and also to find a suitable model for Open Source Software. The phases that are involved in the development process of OSS was given by (Jorgensen, 2001). A comparative study between open and closed software was done by (Ven et al., 2008). In their study, comparison between the two was carried with respect to the development, cost, number of users and software design. Due to the limitations in the time and the resources, it becomes impossible for the developers to deliver a software product meeting the desired level of reliability within a single Software Development Life Cycle. Successive release is planned by the software developers which provides a number of benefits over a single release policy. The benefits that are provided by successive release policy has been studied by (Garmabaki et al., 2012). Kapur et al. (2014) proposed a multi up-gradation model for closed software that considers the issues related to the generation of faults when new functionality is added in the software. The SRGM proposed by them has the capability to capture the faults that are generated due to the add-ons along with the faults from the previous release that have been left undetected.